Convolutional neural network based encoder-decoder architectures for semantic segmentation of plants

Kolhar, Shrikrishna; Jagtap, Jayant

doi:10.1016/j.ecoinf.2021.101373

Cited by 41 publications

(25 citation statements)

References 21 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…According to recent report of Khan et al, 66 for semantic segmentation of prostate in T2W MRI, Deeplabv3 + showed improved segmentation performance as compared to FCN, SegNet, and Unet. In another report on leaf segmentation challenge dataset 67 , the Unet showed better segmentation results as compared to SegNet. Authors reported that the number of trainable parameters in SegNet and Unet were ~ 33.7 and 31 million, respectively.…”

Section: Discussionmentioning

confidence: 94%

A lightweight deep learning model for automatic segmentation and analysis of ophthalmic images

Sharma

Ninomiya

Omodaka

et al. 2022

Sci Rep

View full text Add to dashboard Cite

Detection, diagnosis, and treatment of ophthalmic diseases depend on extraction of information (features and/or their dimensions) from the images. Deep learning (DL) model are crucial for the automation of it. Here, we report on the development of a lightweight DL model, which can precisely segment/detect the required features automatically. The model utilizes dimensionality reduction of image to extract important features, and channel contraction to allow only the required high-level features necessary for reconstruction of segmented feature image. Performance of present model in detection of glaucoma from optical coherence tomography angiography (OCTA) images of retina is high (area under the receiver-operator characteristic curve AUC ~ 0.81). Bland–Altman analysis gave exceptionally low bias (~ 0.00185), and high Pearson’s correlation coefficient (p = 0.9969) between the parameters determined from manual and DL based segmentation. On the same dataset, bias is an order of magnitude higher (~ 0.0694, p = 0.8534) for commercial software. Present model is 10 times lighter than Unet (popular for biomedical image segmentation) and have a better segmentation accuracy and model training reproducibility (based on the analysis of 3670 OCTA images). High dice similarity coefficient (D) for variety of ophthalmic images suggested it’s wider scope in precise segmentation of images even from other fields. Our concept of channel narrowing is not only important for the segmentation problems, but it can also reduce number of parameters significantly in object classification models. Enhanced disease diagnostic accuracy can be achieved for the resource limited devices (such as mobile phone, Nvidia’s Jetson, Raspberry pi) used in self-monitoring, and tele-screening (memory size of trained model ~ 35 MB).

show abstract

Section: Discussionmentioning

confidence: 94%

A lightweight deep learning model for automatic segmentation and analysis of ophthalmic images

Sharma

Ninomiya

Omodaka

et al. 2022

Sci Rep

View full text Add to dashboard Cite

show abstract

“…In the past, numerous researchers implemented Adam optimizer (Bao et al, 2021;Gonzalez-Huitron et al, 2021;Waheed et al, 2020), used a learning rate of 0.001 (Gonzalez-Huitron et al, 2021;Kolhar and Jagtap, 2021;Zabawa et al, 2020), trained the models for one hundred epochs (Abdalla et al, 2019;Kolhar and Jagtap, 2021;Tassis et al, 2021;Zou et al, 2021) and achieved impressive results for agricultural images. Therefore, in the present study, models were trained for one hundred epochs having Adam optimizer (Kingma and Ba, 2017) as an optimization algorithm with a learning rate of 0.001 and a small batch size of ve images as recommended by (Kandel and Castelli, 2020).…”

Section: Training Of Cnn Modelsmentioning

confidence: 99%

“…For the performance evaluation of proposed neural networks and comparisons among these, segmentation results obtained using networks were quanti ed using well-known evaluation metrics for agricultural images: F1-score (Chen et al, 2021;Kestur et al, 2019;Kolhar and Jagtap, 2021), intersection-over-union (IoU) (Sun et al, 2021;Zabawa et al, 2020;Zou et al, 2021), mean intersectionover-union (mIoU) (Long et al, 2015;Tassis et al, 2021;Xu et al, 2020), pixel accuracy (Kestur et al, 2019;Kolhar and Jagtap, 2021;Tassis et al, 2021), precision (Kestur et al, 2019;Zabawa et al, 2020) and recall (Chen et al, 2020;Tassis et al, 2021;Zou et al, 2021) metrics. IoU metric as de ned by Eq.…”

Section: Evaluation Indexesmentioning

confidence: 99%

Lightweight convolutional neural network models for semantic segmentation of in-field cotton bolls

Singh

Tewari

Biswas

et al. 2022

Preprint

View full text Add to dashboard Cite

Introduction Robotic harvesting of cotton bolls will incorporate the benefits of manual picking as well as mechanical harvesting. For robotic harvesting, in-field cotton segmentation with minimal errors is desirable which is a challenging task. In the present study, three lightweight fully convolutional neural network models were developed for the semantic segmentation of in-field cotton bolls. Materials and Methods Model 1 does not include any residual or skip connections, while model 2 consists of residual connections to tackle the vanishing gradient problem and skip connections for feature concatenation. Model 3 along with residual and skip connections, consists of filters of multiple sizes. Effects of filter size and the dropout rate were studied. Results All proposed models segment the cotton bolls successfully with the cotton IoU (intersection-over-union) value of above 88%. The highest cotton IoU of 91.03% was achieved by model 2. The proposed models achieved F1-score and pixel accuracy values greater than 95% and 98%, respectively. The developed models were compared with existing state-of-the-art networks namely VGG19, ResNet18, and InceptionV3. Despite having a limited number of trainable parameters, the proposed models achieved mIoU (mean intersection-over-union) of 93.84%, 94.15%, and 94.65% against the mIoU values of 95.39%, 96.54%, and 96.37% obtained using state-of-the-art networks. The segmentation time for the developed models was reduced up to 52% compared to state-of-the-art networks. Conclusion The developed lightweight models segmented the in-field cotton bolls comparatively faster and with greater accuracy. Hence, developed models can be deployed to cotton harvesting robots for real-time recognition of in-field cotton bolls for harvesting.

show abstract

“…In this study, the training weights achieved by neural networks were implanted into precision robots to remove weeds from field sugar beets. Many scholars have studied plant leaves through semantic segmentation ( Barth et al., 2018 ; Miao et al., 2020 ; Kolhar and Jagtap, 2021 ; Masuda, 2021 ). However, the conventional semantic segmentation network has a large number of parameters and needs a long training time.…”

Section: Introductionmentioning

confidence: 99%

Plant leaf veins coupling feature representation and measurement method based on DeepLabV3+

Liu

et al. 2022

Front. Plant Sci.

View full text Add to dashboard Cite

The plant leaf veins coupling feature representation and measurement method based on DeepLabV3+ is proposed to solve problems of slow segmentation, partial occlusion of leaf veins, and low measurement accuracy of leaf veins parameters. Firstly, to solve the problem of slow segmentation, the lightweight MobileNetV2 is selected as the extraction network for DeepLabV3+. On this basis, the Convex Hull-Scan method is applied to repair leaf veins. Subsequently, a refinement algorithm, Floodfill MorphologyEx Medianblur Morphological Skeleton (F-3MS), is proposed, reducing the burr phenomenon of leaf veins’ skeleton lines. Finally, leaf veins’ related parameters are measured. In this study, mean intersection over union (MIoU) and mean pixel accuracy (mPA) reach 81.50% and 92.89%, respectively, and the average segmentation speed reaches 9.81 frames per second. Furthermore, the network model parameters are compressed by 89.375%, down to 5.813M. Meanwhile, leaf veins’ length and width are measured, yielding an accuracy of 96.3642% and 96.1358%, respectively.

show abstract

Convolutional neural network based encoder-decoder architectures for semantic segmentation of plants

Cited by 41 publications

References 21 publications

A lightweight deep learning model for automatic segmentation and analysis of ophthalmic images

A lightweight deep learning model for automatic segmentation and analysis of ophthalmic images

Lightweight convolutional neural network models for semantic segmentation of in-field cotton bolls

Plant leaf veins coupling feature representation and measurement method based on DeepLabV3+

Contact Info

Product

Resources

About